EQLS 2012 - Methodology

In an effort to provide comparable and reliable data on quality of life across Europe, Eurofound has developed a unified methodological approach and quality assurance system. Over the years, the methodology has been improved, new concerns have been integrated and the geographical scope has expanded with the enlargements of the European Union.

For further information please see the Technical report.

Sampling design

The sample of the EQLS is representative of the adult population living in private households during the fieldwork period in each of the countries covered. In most countries, multi-stage stratified and clustered sampling design was applied.

1. The main sampling stages are as follows: Selection of primary sampling units (PSUs), stratified according to geographic regions (at NUTS2 level or equivalent) and degree of urbanisation. Addresses were clustered in a defined number of PSUs except in the Netherlands, Malta, and Sweden where samples were drawn directly from the registries without clustering.
2. Subsequently, individuals or addresses were randomly selected in each PSU. In 15 countries, up-to-date, high quality address or population registers covering at least 95% of persons or households were available and these were used as a sampling frame. Random route sampling was used in 12 countries where access to register information of sufficient quality was not available., In those countries where the random route method was applied, enumeration of addresses was carried out and validated separately in advance of fieldwork to improve the survey quality.
3. Selection of an interviewee within a household: Once a contact with a household was established, the interviewer followed a defined and documented procedure of respondent selection by firstly registering all adults in the household, and then applying the ‘next birthday rule’ for selecting a person to be interviewed. Only the selected person could be interviewed and only one person per household. The next birthday rule was not necessary in Hungary, Malta, Slovenia and Sweden, where individual respondents were preselected from the name-based registries.

To increase the chance of successful contact and interviewing a randomly sampled person, at least four visits to a preselected address at different times of the day and of the week, spread over at least two weeks, were carried out.

Due to the geographical spread and expected difficulties in achieving collaboration through initial face-to-face contacts, telephone pre-recruitment of respondents was allowed in Finland and Sweden. Only those who agreed on the telephone to participate in the survey were visited by interviewers.

Sample size

The sample size ranges from 1,001 to 3,055 with the target number of interviews set at 1,000 interviews for most countries. The seven EU countries with the largest population, making up altogether three quarters of the EU population, had a higher sample size in order to help to improve the precision of estimates at European as well as national level. The total sample size for all 34 countries was 43,636.

Table: Number of completed interviews

For further information please see the Sampling report.

The EQLS does not include open-ended questions; therefore there was no need for coding this type of question. However, recoding schemes were applied to produce internationally comparable information regarding educational attainment and income. Information about respondents’ educational attainment was recoded into categories of the International Standard Classification of Education (ISCED), while income information that referred to the national currency in each country was recoded to euros based on the exchange rates on 16 May 2011.

Finally, the answer categories for the question on occupation can be matched with the 1 digit ISCO categories that were used in the 5th European Working Conditions Survey which enables comparison between the two surveys for this variable.

For further information please see the Coding report.

The EQLS is a random probability survey of people aged 18 and over living in private households. In order to be representative in terms of gender, age, urbanisation level, region and household size, weighting is required for the EQLS.

Selection probability weights, or design weights (w1)

Adults living in households with more adults in them have a smaller probability of being selected for an interview than people living alone. The selection probability weight correcting this is the number of adults in a household. In the EQLS it is maximized at 4.

Post-stratification weights

Some groups of people are more likely to be available or willing to participate in the survey than others. The post-stratification weights correct for undersampling of certain types of populations, by comparing the EQLS to Eurostat data on adult population by age and gender, urbanization and region, and with EU-SILC data on households by household size.

The final national weight is the product of the selection probability weight and the post-stratification weight.

Cross-national population weights

For results representing the whole EU, weights need to be applied to reflect the size of a country’s adult population relative to the EU adult population to correct for underrepresentation of larger countries. These weights include the selection probability weight and the post-stratification weight.

A Weighting report on the process was published in 2013.

A revision of the weighting strategy for the EQLS took place in 2014 and a new methodology was developed to calculate post-stratification weights. The results of this research are available in the report 'Revision of the weighting strategy in the European Quality of Life Survey (EQLS)', published in 2015.

To ensure high quality of the data, each stage of Eurofound surveys is subject to detailed planning, close monitoring and documentation. Eurofound surveys subscribe to the quality criteria of the European Statistical System (ESS).

In designing the 3^rd EQLS, the results of the quality assessment of the 2^nd EQLS 2007 and the results of a dataset user survey were taken into account, along with consideration of emerging policy issues.

Quality control measures were included in the technical specification of the survey, and a specific Quality control plan was developed and adopted with the fieldwork contractor. This ensured the application of best practice and establishment of specific quality control targets encompassing all survey phases.

The fieldwork contractor has compiled a Quality assurance report which provides a summary of the quality assurance activities performed during the survey preparation, data collection and data processing. Quality control measures for the 3^rd EQLS covered elements from sampling to translation and questionnaire verification to interviewer control and data validation:

• questionnaire verification (pre-tests, pilot interviews);
• questionnaire translation validation;
• sample quality control, including enumeration control (via geocoding and mapping, at least 10% of the PSUs) for countries where registry based sampled were not available;
• checks and approvals of fieldwork materials;
• interviewing verification (CATI, postal or face-to-face back-checks of random 10% of the cases);
• fieldwork visits by Eurofound;
• systematic data validation.

A data editing & cleaning report is also available.

After the completion of the fieldwork, an external quality assessment of the EQLS 2011 was carried out against the quality criteria outlined by the European Statistical System (ESS).